Sei sulla pagina 1di 5

2011The International Conference on Advanced Power System Automation and Protection

Matlab-based voltage stability analysis toolbox and its validity check


WEI YanFang*, WEI ZhiNong, SUN GuoQiang
College of Energy and Electrical Engineering, Hohai University, Nanjing 210098, China;

Abstract: A novel Voltage Stability Toolbox(VST) is introduced, which is programmed using Matlab math language, and the
features of its open source code and a complete set of user-friendly graphical interfaces are detailed. It implements standard
Newton-Raphson (NR) and Newton-Raphson-Seydel (NRS) convergent algorithms to analyze the voltage stability and bifur-
cation of power system. The main functions of VST include power flow, time domain simulation, singularity analysis, eigen-
analysis, as well as static and dynamic bifurcation analysis. VST is a very helpful tool to understand voltage stability and non-
linear bifurcation phenomenon. Test results of the application of this toolbox in a five-bus system with three generators and
30-bus test system are presented to show the practicality and effectiveness of the VST.

Keywords: voltage stability toolbox; Matlab; voltage stability; bifurcation analysis

1 Introduction Software toolboxes for power system analysis can be ba-


sically divided into two classes of tools: commercial soft-
Essentially, the power system is a dynamic nonlinear system; wares and educational/research-aimed softwares. The for-
the property change of voltage stability is a dynamic pro- mer commonly used are PSS/E, ATP, PSCAD etc., those
cess from stable state to bifurcation. Both theory and prac- softwares are typically well-tested and computationally ef-
tice indicate that the bifurcation theory is a powerful tool to ficient. Despite their completeness, these softwares can re-
analyze structure stability mechanism of nonlinear dynamic sult cumbersome for educational and research purposes. For
system. Now bifurcation theory is widely used in power research purposes, the flexibility and the ability of easy
system voltage stability analysis and has become one of the prototyping are often more crucial aspects than computa-
important theory tools in voltage stability research [1-5]. tional efficiency. On the other hand, there is a variety of
Since the resource of solar energy is abundant in China, open source research tools based on Matlab, which are typ-
especially in the west, such as Qinghai Province, Tibet, Xin- ically aimed at a specific aspect of power system analy-
jiang Autonomous Regions, the parabolic trough solar thermal sis[6-7].
power technology has received extensive attention recently. Some of the Matlab-based toolboxes are: Power System
The main motive of this paper is to develop an experi- Analysis Toolbox(PSAT)[6], Matpower, SimPowerSystems
mental platform of the PTCS, and evaluate the performance toolbox(SPS), Power System Toolbox(PST), Voltage Stabil-
of the PTCS. Furthermore, some key parameters that influ- ity Toolbox(VST)[8-10], and so on. Among these toolboxes,
ence the performance of the PTCS, such as the solar flux, VST is open source and freely downloadable. This paper
the flow rate of the HTF, heat loss, are also investigated. introduces the VST, which is mainly used in the analysis for
The encouraging results have been obtained, which can power system voltage stability and bifurcation. The main
provide fundamental data for developing the parabolic function of VST includes: power flow calculation, time do-
trough solar thermal power plant in China. main simulation, bifurcation analysis, singularity analysis,
Comparing with traditional linear analysis approaches, etc.
bifurcation theory is more accurate in analyzing the system This paper first describes the main structure of VST and
stability around critical point neighborhood. But this meth- analysis its main function modules; then a five-bus system
od is against complex calculations and modeling difficul- with three generators and IEEE 30-bus test system are pre-
tiesˈso that appropriate simulation analysis methods are sented to show the applicability of the VST, the simulation
needed to thorough study of power system bifurcation. This results show that the tool is suitable to study the voltage
paper describes the Voltage Stability Toolbox (VST), which stability of power system.
is a flexible tool to analyze power system voltage stability
using bifurcation method.

*Corresponding author (email: meilaier@163.com) APAP2011 www.apap2011.org


___________________________________
978-1-4244-9621-1/11/$26.00 ©2011 IEEE

2011The International Conference on Advanced Power System Automation and Protection

2 Main Function Introduction of VST ìï d = [d ? d ]T


ïï Pg P2 Png
ïï
The VST integrates the symbolic and numeric computations T
í d Pl = [d Png + 1 ? d Png + npq ] (2)
with a graphical menu-driven interface based on MATLAB ïï
and its extended symbolic toolbox. It implements symbolic ïï d = [d T
ïïî Ql Png + 1 ? d Png + npq ]
computations to build exact load flow equations and Jaco-
bian matrices including 2nd–order derivatives, required to where, n is the total node number of the system, generator 1
implement numerical computations such as New- is the swing bus with zero angel and all the other phase an-
ton-Raphson (NR) and Newton-Raphson-Seydel (NRS) for gels are relative to the swing bus; ng and npq are the bus
bifurcation analysis. It has proved to be a useful tool for number of PV and PQ nodes, respectively ˗
education and research in the area of power system stability Pg = [ P2 ? Pn g ]T is pure active power injection vector of
analysis. Figure 1 shows the schematic diagram of the
VST[11]. ng - 1 generators ˗ Pl = [ Png + 1 ? Png + npq ]T and

Ql = [Qng + 1 ? Qng + npq ] are the real and reactive power


T

demands of npq load nodes, respectively; the elements of


d Pg , d Pl and d Ql , can be set to be positive, negative, or
zero depending on the load increase scenario of interest.
Therefore, one may conveniently increase load or gener-
ation at a bus or at a group of buses through equation (1).
VST adopts direct method to obtain static bifurcation
point; the direct method is proposed by Seydel in 1979[12],
which is used to calculate static bifurcation of single pa-
rameter.
Figure 1 Schematic diagram of VST Static bifurcation uses NR-NRS two-stage algorithm to
calculate equalization points. The dynamic bifurcation
analysis identifies small-signal stability characteristics by
Load flow calculation computing and checking the eigenvalues of the reduced
Load flow is the most basic and important calculation of system Jacobian matrix at each equilibrium point along PV
power system analysis. VST implements the standard NR curve. It will occur three kinds of bifurcation along the PV
algorithm to compute the load flow solutions at a given set curve: SNB, Hopf Bifurcation, and Singularity Induced Bi-
of bus injections. The GUI for load flow calculation allows furcation (SIB).
the user to set control variables of the NR algorithm such as
tolerance, error, maximum number of iteration, parameter Small-signal analysis
values and state values. In power system small-signal analysis, system matrix ei-
genvalues have close relations with the system stability.
Bifurcation analysis of voltage stability VST studies small signal stability of a certain point on PV
VST uses a three-stage NR-NRS-NR algorithm to ana- curve by analyzing system matrix eigenvalues.
lyze voltage stability bifurcation; the first stage uses stand- VST analyzes the sensitivity of eigenvalue nearby SIB by
ard NR method to iteration until it fails to converge. Then it the following equations:
automatically switches to the NRS method to find load flow [ J ] = 
solutions at and around the Saddle Node Bifurcation(SNB) (3)
point. After obtaining the SNB point, the standard NR [ g y ] = h 
method is switched back to compute low voltage of PV where, J is the Jacobin matrix of the system, gy is the Jaco-
curve. bin matrix of the algebraic equations, O is the biggest real
Changes in bus injections are achieved through parame- eigenvalue of J, K is the smallest real eigenvalue of gy, H
terization of bus injections with a scalar parameter known and U are the right eigenvector of O and K, respectively.
as a bifurcation parameter[9]:
Dynamic time simulation
m = m0 + a * m' (1)
With the increasing of modern power system scale and
where, P is the initial real or reactive power, D is the bifur-
0

cation parameter which denotes the level of load, operation complexity, power system dynamic stability has
attracted more attention; time domain simulation is an im-
m' = [d TPg d TPl d QTl ]T is the vectors of searching direc-
portant method to research power system dynamic safety
'
tions, T denotes transpose, elements of m are analysis and control. VST also has the function of time do-
main simulation. The time domain simulation adopts an


2011The International Conference on Advanced Power System Automation and Protection

ordinary differential equation (ODE) solver combined with cond-order derivatives. The dynamic bifurcation and singu-
an algebraic solver at each iteration step. A fourth-order lar points of node 4 is shown in Figure 4.
Runga–Kutta method is used as an ODE solver, while an
NR procedure is implemented as an algebraic solver to up-
date the load bus voltage magnitudes and angles at each
integration step.

3 Numerical examples

Simulation and analysis of five-bus system with


three generators
A five-bus system is analyzed firstly, whose connection
diagram is displayed in figure 2 [13-14].

bus 2 bus 4 bus 1


jX d jX d
G G
generator 2 generator 1 Figure 4 Voltage magnitude at bus 4 of dynamic bifurcation and singular
jX d Pd 4ˈQd 4 points
bus 3
jX d
G As shown in Figure 4, the SNB and SIB are identified as
generator 3 the parameter D changing. The singular points and PV curve
bus 5 Pd5ˈQd5
intersect at SIB; as D equal to 3.0205, the system occurs
SIB. Now the stability performance of low voltage equilib-
Figure 2 Five-bus system with three generators rium points has changing from stable to unstable state. The
high voltage and low voltage equilibrium points coincide
1) Dynamic time simulation when D increases to 3.1633.
The voltage amplitude of node 4 is shown in Figure 3; 2) Singularity and eigenvalue analysis
the lower part of Figure 3 is the partial enlarges figure For a point of PV curve, VST analyzes the small-signal
around SNB. As shown in Figure3, when D is 3.1633 the
stability of the point by computing system Jacobin matrix
system occurring SNB, which is the right most of the PV
curve; exceeding this point power flow cannot be converged; eigenvalues. The distributing and movement of two critical
SNB is corresponds to maximum power point at a certain eigenvalues are shown in Figure 5 as D increasing; uncriti-
load increase mode of the system. cal eigenvalues are not figured and the dotted arrow line in
Figure 5 denote the increase direction of D.

Figure 3 Voltage magnitude at bus 4 of static bifurcation

In VST dynamic bifurcation simulation, a fixed-step size


is used at the NR stage, at the NRS stage, the step size is
changed automatically by VST according to the se-


2011The International Conference on Advanced Power System Automation and Protection

Table 2 we can know that the sensitivity of G3 to D is the


biggest, which is 0.6294. The result validates the change
scheme of D.
Simulation and analysis of IEEE 30-bus test system
The searching directions of injection power and injection
space of IEEE 30-bus test system are set. The voltage sta-
bility bifurcation of IEEE 30-bus test system is simulated by
VST, which equally verifies the applicability of VST in an-
alyzing the bifurcation behavior of voltage stability.
Due to limited space, Table 3 and Table 4 list only part
sensitivity parameters of the system state varieties to left
and right eigenvector.

Table 3 Sensitivity of remedial action vector versus D of D of IEEE


30-bus system

State variables P2 P7 Q7 Q11 Q19


0.2072 0.4013 0.1938 0.2089 0.172
Sensitivity of D
2
Figure 5 Critical eigenvalues plot of the system matrix

Just before D equal to 3.0205, the two critical eigen- Table 4 Sensitivity of null space vector versus D of IEEE 30-bus sys-
values are located in the left half plane in Fig.5, which im- tem
plies stability. As the parameter D changing from one of the
State variables G2 V7 G7 G8 G16 G18
complex eigenvalue to the right half plane and becoming a
0.221 0.21 0.39 0.20 0.22 0.20
large positive number while the other eigenvalue stays in Sensitivity of D
7 40 97 87 32 58
the left half plane, but it becomes a large negative real
number. Therefore, stability feature of the equilibrium un- As shown in Table 3 and Table 4, null space vector of V7
dergoes an instantaneous change from stable to unstable. As and G7 are more sensitive to the increasing D. Similarly, re-
D increasing to 3.1633, a critical eigenvalue of the system medial action vector of P7 and Q7 is more sensitive to the
becomes to zero while the other eigenvalue stays in the left increasing D, which means node 7 is the critical node of the
half plane, now the system occurs SNB. system at the load increase mode. The adjacent region of
Table 1 and Table 2 are the sensitivity of the system node 7 is the dangerous region of the system; to analyze the
state varieties to left and right eigenvector corresponding to change of the system variables around node 7 can provide
the zero eigenvalue of load flow Jacobian matrix at the SNB practical information of prevention and remedial control for
point, respectively; where left eigenvector is remedial action voltage stability.
vector and right eigenvector is null space vector. G, V, P, Q
in Table 1 and Table 2 denote phase angle, voltage ampli- 4 Conclusion
tude, active power, reactive power, respectively.
Power system is a complicated nonlinear dynamic system,
Table 1 Sensitivity of remedial action vector versus D of five-bus sys- its operation stability, especially the voltage stability is an
tem important component of modern power system stability
State var-
analysis. VST, a Matlab-based toolbox for voltage stability
P2 P3 P4 Q4 P5 Q5 bifurcation analysis is introduced in this paper, which has
iables
open source code and user-friendly graphical interface. VST
Sensitivity
0.1254 0.5183 0.1762 0.0810 0.4200 0.7083 is very helpful to understand voltage stability and nonlinear
of D
bifurcation phenomena. The feasibility and validity of the
proposed VST are verified by simulating a five-bus system
Table 2 Sensitivity of null space vector versus D of five-bus system with three generators and IEEE 30-bus test system, the re-
sults show that the VST is a powerful tool to analyze volt-
State variables G2 G3 V4 G4 V5 G5 age stability for power system researchers. The new version
0.152 0.62 0.39 0.21 0.60 0.09 3.0 of VST and its download website is available at:
Sensitivity of D
2 94 76 39 60 84 http://power.ece.drexel.edu/vst.html.
As shown in Table 1, the sensitivity of Q5 to D is the
biggest, which is 0.7083. It can be deduced that adjust Q5 This work was supported by the National Natural Science Foundation of
appropriately can further increase the system stability. From China (Grant No. 50877024).


2011The International Conference on Advanced Power System Automation and Protection

1 Kwatny H, Pasrija A, Bahar L. Static bifurcation in electric power 8 Ayasun S, Nwankpa C O, Kwatny H G. Voltage stability toolbox for
networks: loss of steady state stability and voltage collapse. IEEE power system education and research. IEEE Trans on Education,
Trans on CAS, 1986, 33(10): 981-991 2006, 49(4):432-442
2 Lerm A A P, Canizares C A, Silveira e Silva A. Multiparameter bi- 9 Ayasun S, Nwankpa C O, Kwatny H G. Computation of singular and
furcation analysis of the south Brazilian power system. IEEE Trans singularity induced bifurcation points of differential algebraic power
on Power System, 2003, 18(2):737-746 system model. IEEE Trans on Circuits and Systems, 2004,
3 MA Youjie, WEN Hulong, ZHOU Xuesong, et al. Bifurcation analy- 51(8):1525-1538
sis on power system voltage stability. Intelligent Computation Tech- 10 Ayasun S, Dafis C, Nwankpa C O, etc. Symbolic analysis and simu-
nology and Automation, Second International Conference, 2009, lation for power system dynamic performance assessment. Power
3:26-29 Systems Conference and Exposition, IEEE PES, 2006, 1:823-831
4 Ajjarapu V, Lee B. Bifurcation theory and its application to nonlinear 11 Drexel University, Department of Electrical and Computer Engineer-
dynamical phenomena in an electrical power system. IEEE Trans on ing, Voltage Stability Toolbox (VST) - PC/Unix Version, [Online]
Power Systems, 1992, 7(1):312-319 Available at: http://power.ece.drexel.edu/vst.html
5 Kwatny H G, Fischl R F, Nwankpa C O. Local bifurcation in power 12 Seydel R. Numerical computation of branch points in nonlinear equa-
systems: theory, computation, and application. Proceedings of the tions. Numerische MathemaUk, 1979, 33:339-352
IEEE, 1995, 83(11):1456-1483 13 K L Praprost, K A Loparo. An energy function method for determin-
6 Federico Milano. An open source power system analysis toolbox. ing voltage collapse during a power-system transient. IEEE Trans on
IEEE Trans on Power Systems, 2005, 20(3):1199-1206 Circuits System, 1994, 41(10):635-651
7 B Chuco P. Electrical software tools overview. [Online] Available at: 14 Ayasun S. Effects of algebraic singularities on the voltage dynamics
http://eurostag.regimov.net/files/DOC2.pdf of differential-algebraic power system model. European Trans on
Electrical Power, 2008, 18(6):547-561



Potrebbero piacerti anche